EP0091768B1

EP0091768B1 - Measuring characteristics of living tissues by ultrasonic waves

Info

Publication number: EP0091768B1
Application number: EP83301893A
Authority: EP
Inventors: Hirohide Miwa
Original assignee: Fujitsu Ltd
Current assignee: Fujitsu Ltd
Priority date: 1982-04-07
Filing date: 1983-04-05
Publication date: 1989-07-26
Also published as: USRE33672E; EP0091768A2; EP0091768A3; DE3380275D1; US4564019A; JPS58173539A

Description

The present invention relates to the measuring of characteristic parameters of living tissues by transmitting ultrasonic waves into a living body and analyzing reflected waves therefrom.
Conventional systems for obtaining tissue characteristics by analyzing reflected ultrasonic waves of plural frequencies have been proposed by linuma (Japanese Patent "Kokai" No. 38490/74) and Nakagawa (Japanese Patent Publication No. 24798/77). With these systems, however, their operations are based on sound pressure waveforms, so that when the ultrasonic waves have a wide frequency band, like pulses, accurate measurements are impossible under the influence of the phase relationships of respective frequency components, pulse overlapping of continuous reflected waves and phase cancellation in a receiving sensor.
The abovesaid prior art systems can be employed in the case where the living body is composed of several kinds of tissues, an ultrasonic reflector of a definite approximately smooth surface exist at the boundary between adjacent tissues and the reflection factor and the transmission factor of the ultrasonic reflector have no frequency dependence. Such reflection is called specular reflection.
With recent technological progress, however, it has become possible to measure a weak reflection from the tissue between boundaries. In general, the tissue has such a microstructure that cells, capillary vessels, lymphatic vessels, muscular fibers and so forth intertwine complicatedly. A typical size of such a tissue is nearly equal to or smaller than the wavelength of ultrasonic waves. On account of this, reflected waves from the microstructure are accompanied by complex interference owing to phase dispersion and pulse overlapping, introducing in a B-mode tomogram a speckled pattern commonly referred to as "speckle". It has been proven experimentally that reflection from the tissue (backward scattering) has a frequency characteristic, that its reflection coefficient is proportional to the nth power of the frequency, and that the value of n is a characteristic value (a parameter) representing the tissue. It has been reported that n=2.1 to 2.2 in the liver and n=3.3 in the myocardium.
Systems for obtaining the tissue characteristics in such a case have been proposed by Hayakawa and others.

^*1. "Theory of Reflecting Ultrasonic Computer Tomograph Using Plural Frequencies", Japanese Journal of Medical Ultrasonics, vol. 7, Supplement II, 1980.
^*2. "Multifrequency echoscopy for quantitative acoustical characterization of living tissues.", J. Acout. Soc. Am. 69 (6), June 1981, pp. 1838-1840.

Noting the energy value of ultrasonic waves, the system ^*1 conducts a second order differentiation by natural logarithms of their frequencies and a first order differentiation in the direction of depth, by which "a second order differentiated value of the attenuation coefficient of the ultrasonic waves by the natural logarithms of their frequencies" is obtained as a tissue characteristic parameter. According to the system ^*2, energy (or power) values of the ultrasonic waves are obtained through utilization of three frequencies f" f₂ and f₃ and, as a difference value, "the second order differentiated value of the attenuation coefficient of the ultrasonic waves by the natural logarithms of their frequencies" is obtained in the form of a parameter. As experimentally ascertained, it is indicated, when the attenuation coefficient is proportional to the first power of the frequency, the abovesaid parameter is in agreement with an attenuation constant.
The abovesaid Hayakawa system requires complex processing corresponding to the second order differentiation by the natural logarithms of the frequencies, and hence is difficult of real-time processing and poor in SN ratio; further, tissue information on the reflection (backward scattering) is entirely lost. Moreover, the parameters thus obtained are insignificant from the physical viewpoint.
The present invention pertains to the measuring of the frequency dependency of the reflection coefficient and that of the attenuation coefficient of the living tissue separately of each other.
The present invention as defined in Claims 1 and 6, respectively, provides a method and apparatus which can enable separate measurement of tissue characteristic information on the attenuation and reflection coefficients of a living tissue to thereby permit non-invasive measurement of accurate internal living tissue information from the skin surface and which is therefore great utility when employed for a medical diagnosis, a medical checkup for geriatic diseases and so forth.
The present invention can provide a method and an apparatus for obtaining tissue information on not only attenuation but also reflection (backward scattering) coefficients of a living tissue through simple processing which can be executed on the real-time basis, too, and does not introduce much noise.
The present invention can provide a method and an apparatus which obtains energy (or power) values of reflected from a living tissue of ultrasonic waves of at least three frequency bands (or components) and performs arithmetic processing of the energy values to thereby obtain living tissue characteristic values (parameters) of clear physical meaning, such as a frequency slope of an attenuation coefficient of a living tissue and a spatial derivative of a frequency power exponent of its reflection coefficient, through simple processing which can be performed on the real-time basis and does not introduce much noise.
The present invention provides a method in which energies of ultrasonic waves of at least three frequency bands (or components) reflected from a living tissue are obtained, and differences among their logarithms are obtained and then an attenuation slope and/or reflection power exponent derivative of the living tissue are obtained from at least two equations obtained by differentiating the differences in terms of the depth of measurement. Apparatus carrying out the method is also provided.
The "power" mentioned herein is energy per unit time but, in this specification, it is included in the "energy".
Reference is made, by way of example, to the accompanying drawings, in which:-

Figure 1 is a schematic diagram of a three-dimensional model illustrating the manner of actual measurement, explanatory of the principle of the present invention;
Figure 2 is a schematic diagram showing a one-dimensional model obtained by a correction of the model of Figure 1;
Figure 3 is a block diagram illustrating an embodiment of the present invention; and
Figure 4 is a time domain diagram of a received reflected signal.

A description will be given first, with reference to Figure 1, of the principle of the present invention. Figure 1 shows the manner in which an ultrasonic transducer 11 formed by a piezoelectric element, for example, PZT sold by Clevite Inc., held in contact with the skin surface 0 of a living body, transmits thereinto and receives therefrom ultrasonic pulses in a direction Z. Reference numerals 0,1, 2, ... i, i+1, ... and m indicate boundaries of body tissues crossing the Z-axis. Intermediate portions between adjacent boundaries, for instance, 0-1, 1-2, 2-3, ... , i-i+1, ... show, for instance, the outer skin, fat, muscles, ... the liver,... and so forth. Reference numeral 51 designates a focused sound field that is determined by the radius of curvature of an aperture concave of the transducer 11 and the frequency of the ultrasonic waves used.
The intensity of received waves reflected from a depth z varies as a function of the depth z with a factor determined by three-dimensional geometric conditions which are dependent on the convergence of a beam on both ways (in both directions) and the wavelength, such as the degree of focusing, even if the subject under test is not a living body but water or the like which does not attenuate ultrasonic waves. The ratio of the reflected waves at the depth z=0 is defined as G_l(z). This ratio can be measured by disposing a perfect reflector, such as a metal plate, at the depth z in water or the like.
Transmitted ultrasonic waves having reached the boundary i are reflected back or scattered (reflected, refracted) and aside three-dimensionally since the living tissues on both sides of the boundary i have different acoustic impedances or the boundary i has irregularities; however, since the acoustic impedance and the speed of sound in the living body do not widely differ with tissues, the transmitted ultrasonic waves mostly pass through the boundary i. Letting the transmission factor, reflection factor and scattering factor of the boundary i with respect to the incident waves thereon be represented by τi, yi and 6i, respectively, they bear the following relationships:
Therefore, even if yi and 6i somewhat have frequency dependence, _Ti can be regarded as having no frequency dependence.
Through using corrections by the abovesaid G_i(z) and _Ti, such a three-dimensional model as shown in Figure 1 can be converted into such a one-dimensional model as shown in Figure 2. The following description will be given of the one-dimensional model. In Figure 2 the transducer has indefinite expanses in the X and Y directions vertical to the ultrasonic transmission direction Z.
The voluminal tissue sandwiched between the boundaries i and i+1 has a microstructure of a typical size nearly equal to or smaller than the wavelength of the ultrasonic waves and having cells, capillary vessels, fibers, nerves and so forth intertwined complicatedly. The microstructure cannot theoretically be measured because of its size relative to the wavelength of the ultrasonic waves and only a spatial mean value of the microstructure can be measured. It has been proven experimentally that ultrasonic waves transmitted into a living body are subjected to attenuation which varies exponentially as the depth (z) increases, and that its attenuation constant a is proportional to the frequency f of the ultrasonic waves. That is,
where β is a proportional constant. The proportional constant β is a frequency slope of the attenuation constant and referred to as an attenuation slope and it is a characteristic value of the tissue.
A reflection from a voluminal tissue is statistically a speckled reflection commonly referred to as "speckle" and a reflection coefficient y is given experimentally as follows:
where b and n are constants. The constant n is a frequency power exponent of the reflection coefficient and a characteristic value of the tissue.
In Figure 2, let it be assumed that ultrasonic waves of a frequency f, (frequency band 20, where 0 is a half-width, and amplitude Q,(O)) are transmitted and received. Alternatively, it may also be assumed that pulses of a wide frequency band are transmitted, the amplitude of a component of the frequency f₁ is Q₁(0) and the component of the frequency f, of reflected waves is observed. Letting the attenuation compensating time control amplification degree (TGC) of the received signal and apparent energy observed including the attenuation compensating time control amplification degree (TGC) be represented by A₁ (z) and E₁ (z), respectively, E₁ (z) is given by the following equation. For convenience of description, the following expression is a computational expression obtained from sound pressure measurement but, if the power of the waves can be directly measured, a direct representing expression can be used.
where F(f₁ · z) is a correcting term obtained by integrating frequency dependent components of reflection and attenuation in connection with a frequency variation from f₁ in the band 20. When Ω<<f₁, F(f₁, Z) can be regarded as invariable at the frequency f,. _T'i is the transmission factor of the reflected waves at the boundary i. n indicates multiplications of TiT'i from i=O to i=i. The upper limit i is a maximum number from the skin surface to the depth z.
Apparent energy E₂ similarly observed in respect of a sound frequency f₂ is as follows:
when Ω₁<<f₁ and Ω₂<<f₂, it can be regarded that F(f₁, z)=F(f₂, z), where Ω₁, and Ω₂ are half-widths.
A difference between natural logarithms of Eqs. (1) and (2) is as follows:
Differentiating (or differencing) Eq. (3) with respect to the depth z (in a reflected signal train, z=ct, where c is the sound speed, and Eq. (3) may also be differentiated with respect to time t.), it follows that
Similarly setting up the equation corresponding to Eq. (4) with respect to f₂ and f₃ by introducing a third frequency f₃, it follows that
In Eqs. (4) and (5), E₁, E₂ and E₃ are measured as functions of the depth z.
Accordingly, the left sides of Eqs. (4) and (5) are both measured values.
The left sides of Eqs. (4) and (5) can be transformed as follows: and
where
corresponds to true energy obtained by correcting apparent observed energy E,.
On the right sides of Eqs. (4) and (5),
and β(z) are unknown and f₁, f₂ and f₃ in that coefficient are known. Accordingly, from simultaneous algebraic equations with two unknowns of Eqs. (4) and (5)
β(z): frequency slope of attenuation coefficient, and
spatial derivative of frequency power exponent of reflection coefficient

can be obtained as functions of the depth z.

The above is a description of the principle of the present invention. The present invention is free from the influence of the boundary transmission factors τi and _T'i, the constant b(z) of the reflection coefficient and the absolute values Q₁, Q₂ and Q₃of the amplitudes of respective frequency components. While in the foregoing Eq. (5) is set up using f₂ and f₃, it may also be set up using f₃ and f₄. Furthermore, if the value A of the attenuation compensating time control amplification degree (TGC) is held constant with respect to all frequencies, then
become zero, making compensation unnecessary.
Eq. (1) and so on can be represented using the amplitude absolute value Q,(z) of the sound pressure but, in order to avoid a bad influence of the phase term, it is necessary to obtain E, first and then Q, from
This is not so significant, and hence is not described in this specification.
By putting the three frequencies f₁, f₂ and f₃ into a geometric progression, the subsequent calculations are simplified. Letting
the right sides of Eqs. (4) and (5) respectively become as follows:
Furthermore, obtaining a difference between Eqs. (4) and (5), the term
is eliminated as will be seen from Eqs. (8) and (9). That is,
Thus the calculation for obtaining β(z) is simplified.
Also the calculation is simplified by putting the frequencies f₁, f₂ and f₃ into an arithmetic progression, that is,
In this case, when obtaining the difference between Eqs. (4) and (5), (z) is eliminated and
can be obtained easily.
Since the pulse length of reflected waves from the depth z usually has a finite value that is not 0, the reflected waves are superimposed on reflected waves from irregular tissues before and after the depth z and certain frequency components may sometimes be irregularly added to or subtracted from each other due to interference, leading to a noticeable error. This is called spectrum scalloping. To avoid this, it is preferable that N values of (3(z) and
obtained by the above-described method are subjected to statistical processing through utilization of as many combinations of frequency components as possible in a utilizable frequency band, for example, N sets of frequency components fj_i, fj₂ and fj₃ (where j=0 to N), thereby obtaining their mean values.
It is needless to say that the calculations for obtaining a(z) and
can be simplified by using the frequencies in the form of a geometric or arithmetic progression.
For improving the statistic precision, it is also possible to transmit and receive pulses and measure them M times for the same scanning line and perform statistical processing of the measured values, various quantities during computation and last computed values. For instance, even if an examinee stops his breath during measurement, tissues on the scanning line wobble three-dimensionally owing to pulsation of the heart, small changes in his posture and so forth. Accordingly, observed values at the depth z in the M-times measurement aredistributed in a specific space domain around a mean center X · Y · Z and the M-times statistical processing bears the same meaning as statistical processing of measured values at M measuring points in a certain space domain. By conducting the statistical processing in connection with L points before and after the depth z for each scanning, the statistical accuracy is further improved.
By this, a maximum of LxMxN samples can be obtained.
If the tissue characteristic value thus obtained as a function of the depth z on a certain scanning line is displayed on the corresponding scanning line on a CRT or the like as in the case of a B-mode graph, a two-dimensionally or three-dimensionally distributed image can be obtained. This is useful for finding out an abnormality, such as a cancer or the like, by visual examination.
A description will be given of the method of the present invention. Figure 3 illustrates an example of apparatus suitable for carrying the present invention into practice. In Figure 3 reference numeral 11 indicates a wide-band transducer, which is formed by piezoelectric elements of the aforementioned PZT or PVDF (polyvinylidene fluoride sold by Kureha Kogyo of Japan). The transducer 11 is shown to be a compound transducer which comprises a PZT transducer 11' and PVDF transducer 11" covering different frequency bands. It is also possible to employ such a system in which the transducer is formed by three layers of center frequencies f₁, f₂ and f₃ for covering frequency bands 2Ω₁, 20₂ and 20₃ (where Ω₁, Ω2 and Ω₃ are half-widths), respectively, and received waves are separated by filters to obtain energies E₁, E₂ and E₃. In this illustrated example, however, pulses of wide band are transmitted and the DFFT (Digital Fast Fourier Transformation) is used. Reference numeral 12 designates a driver, which may be arranged to drive the transducers 11' and 11" by impulses or in separately specified frequency bands. Reference numeral 13 identifies a wide-band amplifier for amplifying signals of received reflected waves. The received signal varies with time as shown in Figure 4 with the time of transmission of pulses represented by t=0, and signals are reflected back from deeper positions in a living body with the lapse of time. The reflected wave from a depth zi appears at such a time as follows:
where C is the sound speed in the living body. A signal from a tissue between the depth zi and zi+Az appears in the following time interval:
Accordingly, the tissue characteristic between the depths zi and zi+Az can be obtained by analyzing the signal received in the time interval Δt.
Since the reflected signal decreases exponentially with an increase in the depth z by virtue of attenuation on forward and backward paths, the amplification degree Z of the amplifier 13 is varied with an increase in the depth z or with the lapse of time t. This is called time-gain control or sensitivity-time control. This control is needed for retaining excellent SN ratio in subsequent signal processing.
Reference numeral 14 denotes a gate, which is opened in the time phase of Eq. (12) and closed after the lapse of time given by Eq. (13). Reference numeral 15 represents an AID converter, which is required to have a sampling speed of about 20 MHz for signals of 1 to 10 MHz band. Assuming that Δz=1.5 mm, then C=1500 m/s, so that Δt=2 µs and, if sampling is carried out at 20 MHz, about 100 samples can be obtained.
Reference numeral 16 shows a DFFT (Digital Fast Fourier Transformation) circuit, which analyzes the abovesaid 100 data to output real parts and imaginary parts of about 50 frequency components. For instance, in the case of the frequency f₁, a component in-phase with cos2πf₁t is a real part R₁ and a component in-phase with sin2πf₁t is an imaginary part l₁.
Reference numerals 17-1, 17-2,... refer to calculating units, which are supplied with the real parts and the imaginary parts of components of the frequencies f₁, f₂, f₃, f₄, ... from the DFFT circuit 16. Since the calculating units 17-1, 17-2, ... are identical in construction and in operation, a detailed description will be given of the calculating unit 17-1 alone.
The calculating unit 17-1 receives the real part R₁ and the imaginary part l₁, ofthe frequency f₁ from the DFFT circuit 16. The real part R₁ and the imaginary part l₁, are squared by square circuits 171 and 172 to obtain R,² and l₁ ², which are added by an adder 173, obtaining the sum R₁ ²+l₁ ². This sum is equal to E₁. Next, InE, is obtained by a logarithmic amplifier 74. Further, InG₁ ²· A₁ ² is obtained as a function of the depth z (or the time t) and prestored in the form of a table in an ROM 178, from which is read out a value for the corresponding z (or t). The output InE, of the lorarithmic amplifier 174 and the output InG₁ ²· A₁ ² of the ROM 178 are applied to a subtractor 175, wherein a subtraction InE₁-InG₁ ²· A₁ ² is carried out to output
which is stored in a memory 176.
Similar processing is performed for the reflected signal received at a time ti+1 after At to obtain
and a difference between this and
at the time ti stored in the memory 176 is obtained by a subtractor 177. The difference thus obtained is a differentiated (differenced) value at Δz. This becomes the output of the calculating unit 17-1 and expresses the following equation:
Likewise, the calculating unit 17-2 provides the following output:
Reference numerals 18-1, 18-2, 18-3, ... signify subtractors. The subtractor 18-1 subtracts the output of the calculating unit 17-2 from the output of the calculating unit 17-1. The subtractor 18-2 subtracts the output of the calculating unit 17-3 from the output of the calculating unit 17-2. In a similar manner, the following subtractors operate.
Thus the output of the subtractor 18-1 provides the difference between Eqs. (14) and (15):
This is the left side of Eq. (4) as shown in Eq. (6). The order of calculation by the calculating units 17-1 and 17-2 and the calculation by the subtractor 18-1 is reverse from the order of calculations described previously but, in this case, it does not matter mathematically.
The output of the subtractor 18-2 similarly provides the left side of Eq. (5).
Reference numeral 19-1 indicates an algebraic calculator which receives the outputs of the subtractors 18-1 and 18-2 and solves from Eqs. (4) and (5) a simultaneous equation with β(z) and
unknown. Certain constants α₁₁ and a_l2 determined by the frequencies f₁, and f₂ are multiplied by the outputs of the subtractors 18-1 and 18-2 and then added together to obtain β(z). And different constants a₂, and α₂₂ are likewise multiplied by the outputs of the subtractors 18-1 and 18-2 and then added together to obtain -
It is convenient to calculate the constants α₁₁, α₁₂, α₂₁ and a₂₂ from the frequencies f₁ and f₂ in advance and prestore them in the algebraic calculator 19-1.
Reference numeral 20 designates an arithmetic mean circuit which comprises an adder 21 for adding the outputs of the algebraic calculators 19-1, 19-2, ... and a divider 22 for dividing the output of the adder 21 by the number N of inputs to the adder 22. The arithmetic mean circuit 20 obtains an arithmetic means value of the N values β(z) or
sequentially obtained for each frequency component of the output of the DFFT circuit 16.
Reference numeral 23 identifies a shift register which comprises L stacked registers 23-1, 23-2,... 23-L for storing the output of the arithmetic mean circuit 20. At first, the output of the arithmetic mean circuit 20 for the depth zi is written into the register 23-1 and when the output of the arithmetic mean circuit 20 for the next depth zi+1 goes into the register 23-1, the content of the register 23-1 is shifted to the register 23-2. In this way, upon each occurrence of inputting new data into the register 23-1, previous data are shifted upward through successive registers in the shift register 23. In consequence, L data are stored in the shift register 23, with the oldest data in the register 23-L and the latest one in the register 23-1.
Reference numeral 24 denotes an arithmetic mean circuit for obtaining an arithmetic mean value of L data. The arithmetic mean circuit 24 is also comprised of an adder 25 for adding L outputs from the registers 23-1 to 23-L and a divider 26 for dividing the output of the adder 25 by L. The outputs of the registers 23-1 to 23-L are added together by the adder 25 and its output is applied to the divider 26, wherein it is divided by L to obtain the arithmetic mean.
The output of the arithmetic mean circuit 24 provides, for each scanning, a mean value of (LxN) β(z)'s or
for each of the depths z₁, z₂, _Z3, ... z and z₁₊₁, and the mean value is stored in a memory. By scanning the same tissue M times at certain time intervals, obtaining a measured value for each scanning, storing it and averaging the values for the same depth z in all the measurements, it is possible to obtain a mean value of LxMxN samples for each depth z,.
While the above description has been given of a method for executing statistical processing with the last calculated value β(z) or
the statistical processing can be applied to intermediate results and this may sometimes make the subsequent calculations easy. This can be achieved, for example, by executing statistical processing of the outputs of the DFFT circuit 16 in connection with frequency for M-time scanning of L points to remove the influence of the spectrum scalloping and executing again statistical processing with a last calculated value.
In the foregoing embodiment the frequency components f₁, f₂, f₃, ... correspond to the outputs of the DFFT circuit 16 in a sequential order but, by a suitable selection of the outputs of the DFFT circuit 16 in a manner to form a geometric or arithmetic progression as described previously, the calculating circuits of the algebraic calculators 19-1, 19-2, ... can be simplified although the number N decreases.
By scanning one sectional area of a living body in successive scanning directions so that, for instance, β(z) may be obtained as a function of each of the depths z₁, z₂, ... z and z₁₊₁ as a mean value of the LxMxN measured values for each scanning direction, and then displaying the resulting values on the corresponding scanning lines of a CRT, it is possible to obtain a distribution diagram of j3(z) or
on the sectional area of the living body. This is very useful for detecting an abnormal tissue as of a cancer.

Claims

1. A method for measuring ultrasonic characteristics of living tissue, comprising:

(a) transmitting, into a living body comprising said tissue, ultrasonic sound pressure waves comprising at least three independent bands centered at frequencies f₁, f₂ and f₃, respectively, or a wide band including such centre frequencies;

(b) obtaining energies E₁, E₂, E₃ or amplitude absolute values Q₁, Q₂, Q₃ of the reflected waves received in each of said bands, employing a receiving sensor (11) arranged to provide an output proportional to the sound pressure of the reflected waves, or a receiving sensor (11) arranged to provide an output proportional to the energy of reflected waves;

(c) taking logarithms InE₁, InE₂ and InE₃ or InQ₁, InQ₂ and InQ₃ of the energies orthe amplitude absolute values, respectively;

(d) obtaining a difference between the two logarithms of a first pair selected from the three logarithms;

(e) obtaining a difference between the two logarithms of a second pair selected from the three logarithms, different from the first pair; and characterised by

(f) determining the derivative of each of the differences with respect to depth into the tissue in the direction of transmission of the ultrasonic waves; and

(g) obtaining from said derivatives the frequency slope of the attenuation coefficient and the spatial derivative of the frequency power exponent of the reflection coefficient of the tissue.

2. A method according to claim 1 wherein the at least three frequencies f₁, f₂ and f₃ form a geometric progression, i.e. f₁/f₂=f₂/f₃=..., and wherein the difference between the two derivatives is formed to obtain the frequency slope of the attenuation coefficient.

3. A method according to claim 1 wherein the at least three frequencies f_i, f2 and f₃ form an arithmetic progression, i.e. f,-f₂=f₂-f₃=... and wherein the difference between the two derivatives is formed to obtain the spatial derivative of the frequency power exponent of the reflection coefficient.

4. A method according to any one of claims 1 to 3 wherein more than three frequencies are transmitted; energies or amplitude values being obtained in respect of N sets of three frequencies of the reflected waves, the measured results being subjected to statistical processing.

5. A method according to any one of claims 1 to 4 wherein the measurement of the energies is made at a specified depth or/and M adjoining depths in the same direction of transmission and reception of the ultrasonic waves; the M measured values being subjected to statistical processing in steps (b) to (g).

6. Apparatus for measuring ultrasonic charactristics of living tissue, the apparatus comprising:

(a) means (11,12) for transmitting, into a living body comprising said tissue, ultrasonic sound pressure waves comprising at least three independent bands centered at frequencies f_i, f₂ and f₃, respectively, or a wide band including such centre frequencies;

(b) means (16, 171 to 173) for obtaining energies E_i, E₂, E₃ or amplitude absolute values Q_i, Q₂, Q₃ of the reflected waves received in each of said bands, said means comprising a receiving sensor (11) arranged to provide an output proportional to the sound pressure of the reflected waves, or a receiving sensor (11) arranged to provide an output proportional to the energy of reflected waves;

(c) means (174) for taking logarithms InE,, InE₂ and InE₃ or InQ,, InQ₂ and InQ₃ of the energies or the amplitude absolute values, respectively;

(d) means (18-1) for obtaining a difference between the two logarithms of a first pair selected from the three logarithms;

(e) means (18-2) for obtaining a difference between the two logarithms of a second pair selected from the three logarithms, different from the first pair; the apparatus being characterised by

(f) means (176,177) for determining the derivative of each of the differences with respect to depth into the tissue in the direction of transmission of the ultrasonic waves; and

(g) processing means (19-1, 19-2) for obtaining from said derivatives the frequency slope of the attenuation coefficient and the spatial derivative of the frequency power exponent of the reflection coefficient of the tissue.

7. Apparatus according to claim 6 wherein the at least three frequencies f,, f₂ and f₃ form a geometric progression, i.e. f₁/f₂=f₂/f₃=..., and wherein the processing means (19-1, 19-2) is arranged to form the difference between the two derivatives to obtain the frequency slope of the attenuation coefficient.

8. Apparatus according to claim 6 wherein the at least three frequencies f,, f₂ and f₃ form an arithmetic progression, i.e. f₁-f₂=f₂-f₃=... and wherein the processing means (19-1, 19-2) is arranged to form the difference between the two derivatives to obtain the spatial derivative of the frequency power exponent of the reflection coefficient.

9. Apparatus according to any one of claims 6 to 8 wherein more than three frequencies are transmitted; the means (16,172,173) for obtaining energies or amplitude values are arranged to perform a measurement of N sets of three frequencies of the reflected waves, the measured results being subjected to statistical processing.

10. Apparatus according to any one of claims 6 to 9 wherein the measurement of the energies is made at a specified depth or/and M adjoining depths in the same direction of transmission and reception of the ultrasonic waves; the M measured values being subjected to statistical processing in means (b) to (g).